W3C home > Mailing lists > Public > public-cwm-talk@w3.org > January to March 2010

Re: scope of _:existentials in N3Logic?

From: <jos.deroo@agfa.com>
Date: Mon, 1 Mar 2010 22:36:32 +0100
To: yves.raimond@gmail.com
Cc: connolly@w3.org, Graham Klyne <GK-lists@ninebynine.org>, public-cwm-talk@w3.org
Message-ID: <OF06F82202.CAB57557-ONC12576D9.0073EC80-C12576D9.0076B387@agfa.com>
Hi Yves,

The euler.yap code is tuned for Yap and for a while we had a clone for SWI
namely http://eulersharp.sourceforge.net/2006/02swap/eyes.pl
When comparing eyes.pl performance with
http://eulersharp.sourceforge.net/2003/03swap/dtb-2010.txt
the numbers are

   10     0.000 sec
  100     0.030 sec
 1000     1.920 sec
10000   435.501 sec

hence about 1000 times slower than eye for depth 10000 and I guess
more than 1 million times slower than eye for depth 1 million.
This is IMHO mainly due to yap's really nifty demand driven indexing
http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.63.6058&rep=rep1&type=pdf

Today Yap is also moving closer and closer to SWI compat
http://sourceforge.net/mailarchive/message.php?msg_name=f710c6c81003010124x3d17649cgc76650469ec220c6%40mail.gmail.com

That said, the scope of _:existentials in N3Logic is indeed worthwile to 
be
tested independently and thanks for your plan to have a deeper look :-)

Kind regards,

Jos De Roo | Agfa HealthCare
Senior Researcher | HE/Advanced Clinical Applications Research
T  +32 3444 7618
http://www.agfa.com/w3c/jdroo/

Quadrat NV, Kortrijksesteenweg 157, 9830 Sint-Martens-Latem, Belgium
http://www.agfa.com/healthcare



Yves Raimond <yves.raimond@gmail.com> 
03/01/2010 07:10 PM

To
Jos De Roo/AMDUS/AGFA@AGFA
cc
connolly@w3.org, Graham Klyne <GK-lists@ninebynine.org>, 
public-cwm-talk@w3.org
Subject
Re: scope of _:existentials in N3Logic?







Hello Jos!

Hi Yves, 

Let me first say that I really like your Henry work esp. the N3 DCG. 
Also thanks to Graham for sharing his Swish experience. 

As you say, in Euler we can use Skolem functions via the e:tuple builtin 
but in the latest Euler yap engine we don't need the Skolem functions 
anymore, we use a "Skolem machine" and "labelvars" instead, 
see the predicates sem/1, sem/4 and labelvars/4 in the code at 
http://eulersharp.sourceforge.net/2006/02swap/euler.yap 


Wow - at first glance that looks amazing! I really like the idea of using 
functors for variable quantification - I will definitely take a deeper 
look into it! Did you try that code with SWI already, or does it only run 
on Yap?

Many thanks,
y
Received on Monday, 1 March 2010 21:37:07 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 1 March 2010 21:37:15 GMT